AD FIELD NOTESAI advertising · the work behind the output

Case files / ai voice interface

Sherwin-Williams Speaking in Colour: make a search system answer a sensory brief

Speaking in Colour translated spoken descriptions into personalised colour palettes through image search, optical recognition and a voice-controlled interface.

Campaign period: 2022Source published: Date not statedArchive entry prepared: 15 September 2026

Retrospective coverage, prepared in September 2026 from the original record.

Speaking in Colour campaign interface with an audio button symbol and colour palette.
WPP campaign image for Sherwin-Williams’ Speaking in Colour. Sherwin-Williams Coil Coatings / Wunderman Thompson; image published by WPP.

Reported execution: speak the colour, then refine it

WPP describes Speaking in Colour as a voice-controlled tool made for Sherwin-Williams Coil Coatings. A user says a word or phrase; the system analyses millions of images through a search algorithm and optical recognition, then produces a palette. The user can add language such as ‘more turquoise’ or ‘dappled sun’ to change the result.

The campaign’s AI role is specific and legible. It does not pretend a machine has discovered a universal emotional meaning for colour; it turns a subjective verbal prompt into an editable starting point.

Our reading: ambiguity is the product brief

Colour selection often starts with language that is evocative rather than technical. The useful interface move is to accept that ambiguity and return something people can react to. A palette is not the answer; it is a conversation object between an imagined place, a user and a material decision.

That only works if revision remains easy. A system that treats its first interpretation as authoritative turns a creative aid into a confusing gatekeeper. Keep the prompt visible, offer adjustments and preserve a human colour review for production use.

Evidence boundary: personalised output is not universal perception

WPP establishes the intended mechanism and gives examples of the interactions. The page does not establish that the generated palette accurately represents every user’s description or that it improves a specified business outcome.

The transferable lesson is practical: use AI to make a fuzzy prompt manipulable, then make clear that the user, not the model, owns the final colour decision. Accessibility testing must remain part of that decision. Teams should also test whether the interface recognizes different accents, speech patterns and descriptive traditions before presenting it as universally personal.

Sources & limits

WPP agency case study; its claims about scale and cultural insights are not independently audited.

  1. WPP — Sherwin-Williams Coil Coatings’ Speaking in Colour — source publication date not stated.

Send a correction with the passage and supporting source.

The weekly field note

One useful note for the next brief.

Campaigns worth studying, the work behind them, and questions to bring into your next creative review. Delivered weekly.

How we handle your email